PropertyValue
is nif:broaderContext of
nif:broaderContext
is schema:hasPart of
schema:isPartOf
nif:isString
  • For the following experiments, the stimuli used were digital photographs of scenes filtered to contain either low (LSF) or high spatial-frequency (HSF) information only. All scene stimuli were created from photographs obtained collectively by the lab. Images contained indoor environments, cityscapes, or landscapes, without any conspicuous human characters or animals. All images were converted to greyscale, and resized to 1000×750 pixel images using Matlab (Mathworks, Natick, MA). They were filtered using a Gaussian filter, with a cut-off frequency of 2 cpd (cycles per degree) for low spatial-frequency images (keeping all frequencies below this value), and 6 cpd for high spatial-frequency images (keeping all frequencies above this value). These cut-off values are typical for filtering images [31], [40], and provide a distance and image-size independent measure of spatial frequency. Luminance values were tested with a customized Matlab protocol, which used saturation values in the red, green and blue channels to estimate luminance. This step was implemented to ensure that the behavioural effects relating to the spatial frequency filtering were not overshadowed by other low-level differences of perceptual saliency in the images, which result from the filtering process itself. Luminance values of the filtered scenes were extracted and tested for differences with two-sample independent t-tests, which were found to be non-significant (comparing non-filtered to LSF, t(286) = –.52; p = .61; non-filtered to HSF, t(286) = –.06; p = .95; HSF to LSF, t(286) = .33, p = .74). All participants were volunteers recruited from a subject pool at the University of Oxford, and gave written consent to participate in this study for monetary compensation. The studies were approved by the University of Oxford Central University Research Ethics Committee (CUREC).
rdf:type